Genetic algorithms

The principle methods in finding good solutions to discrete optimisation problems are discussed. Genetic Algorithms (GAs) are methods which start with a population of trial solutions, and by processes which mimic those found in evolution and natural selection, the population is improved from generation to generation until a population is arrived at containing one or more good solutions. When a good solution spontaneously arises in one individual, then this tends to move the entire population of solutions towards that region of the solution space. Thus, when the solutions migrate to the region of the global optimum, the coverage of the solution space is more complete and there is less likelihood of becoming stuck in a local minimum. It is this feature that makes GAs of such interest, and therefore the subject of this paper. The aim of this paper is to give a brief description of the algorithm, followed by some implementation considerations. Then, some existing and possible applications to digital signal processing, circuit design and related topics are presented, together with possible research directions